ISSUE 399  ·   August 9, 2022

Insight

Analysts Should Have Portfolios

Stop me if this sounds familiar: “I can’t get a job without experience, but I can’t get experience without a job.” Demonstrating your abilities is tough, particularly doing so before your first interview. Benn Stancil suggests portfolios as a better approach, and I think he’s onto something.
Benn Stancil

What would a theory of data visualization look like?

What makes one data visualization more effective than another? This post from Enrico Bertini examines this question and lays out a working theory for how to best communicate as much quality information as possible.
FILWD | Enrico Bertini

Sponsored Link

Deploy Any ML Model to Any Cloud Platform

Are your model deployments an infrastructure nightmare? Use Truss, the open-source library for model packaging and deployment, developed by Baseten. Truss bridges the gap between model development and model deployment by making it equally straightforward to serve a model on localhost and in prod, making development and testing loops rapid.

Tutorials, Projects & Opinions

Introduction to Streaming for Data Scientists

As machine learning moves towards real-time, streaming technology is becoming increasingly important for data scientists to understand. This in-depth post from Chip Huyen covers where streaming is useful, why it’s hard, and how it can help you.
Chip Huyen

Unchecked A/B Testing Destroys Everything it Touches

“Every infuriating thing on the web was once a successful experiment” is an interesting premise for a post. As data scientists, our focus is often on making sure experiments are statistically sound, but it’s also important to keep product context in mind and push back when needed.
Derek Zumsteg

The Random Forest Algorithm

Machine Learning University, an education initiative from Amazon designed to teach machine learning theory and practical application, covers random forests in a great visual explainer of how things work behind the scenes.
MLU Explain | Jenny Yeon & Jared Wilber

Interim Analysis & Group Sequential Designs

One common knock on A/B testing is that it takes too long. Depending on your sample size and experimental design, tests can take weeks or even months to get a definitive result. Peeking is way to speed things up, but it comes with caveats. This post explores ways to peek while preserving type one error rate.
Demetri Pananos

SQL: The Universal Solvent for REST APIs

I enjoyed this forwarding-looking post on what accessing data from REST APIs could look like once more of the low-level work has been abstracted away. Being able to jump directly to writing SQL seems like a win to me.
O'Reilly Radar | Jon Udell

Twitter Spaces: Responsible and Trustworthy AI

Listen in on a live Twitter Spaces at 9:30am PST on Wednesday, August 17, featuring Ben Lorica of Gradient Flow, Andrew Burt of bnh.ai, and Bob Friday of Juniper Networks. The conversation will discuss what it takes to get to mature adoption of AI in the enterprise, and the role trust plays throughout. 
// sponsored

How did you like this issue of Data Elixir?

👎    1     2     3     4     5   👍

Was this email forwarded to you?  Sign up here >>